While I mostly agree with you that upgradeability and repairability are crucial, and have done my share of rescuing dumpster-dived laptops and refitting them for use, IMO this approach we're seeing here is a sign of what needs to happen. For the types of laptop designs that are popular - thin-and-lights in the 12-14" class - dual SO-DIMM slots is ... well, questionably feasible at best. CPUs have AFAIK been soldered since Ivy Bridge (or was it Haswell?), which is a shame, but given the reduction in Z-height for LGA vs the previous PGA sockets, that's never being reversed. That's just a fact of life, there just isn't room in a modern ~15mm thick laptop. There's no excuse for soldered storage though - any device can fit an m.2 2240 SSD (though there's definitely an argument for making a lower profile connector for this as well).
The Framework laptop manages impressive modularity for its size and featureset, but also has a relatively small battery and thus mediocre battery life due to this. Of course other laptops have the same bettery size and life due to just being stupidly thin, which is another problem entirely. But looking inside the Framework, the area taken up by the dual SO-DIMMs is huge. If that could be reduced by even 1/4, that would be very welcome, especially if this also comes with a lower z-height for the memory module and connector. If gains like this can be had by moving to a module that integrates dual channel RAM into a single module, I'd say that's a decent tradeoff overall, even if it rules out the "buy xGB, single-DIMM laptop, add same capacity DIMM for 2xGB laptop on the cheap" approach to upgrading. And, of course, it would rule out the bane of budget laptops: the single SO-DIMM, single-channel design. That alone would be worth the move to a new standard to me.
IMO, to safeguard upgradeability and repairability for laptops we need new, future-oriented connector standards that are suited to the products people actually want (as well as manufacturers willing to sell motherboards for CPU/platform upgrades, though IMO that's rather unlikely outside of niche outfits like Framework). We need to maintain the importance of repairability and upgradeability, but we cant tie those demands to a demand for unsuitable, old connectors. The sockets we have now work well for what they do, but none of them are designed for the density required in contemporary designs. Some quick googling tells me that a standard DDR3 SO-DIMM socket is 5.2mm tall (tall versions are 9.2mm), and DDR4 sockets are quite similar. If that could be cut to just the thickness of the PCB+memory chips (and VRMs for DDR5) - I would guess < 3mm), that would make adding upgradeable memory a lot easier on thin-and-light designs, especially if the socket/connector could be mounted to a PCB edge with the memory module overhanging its edge. In a 15mm-ish laptop, another 2-ish mm from lower profile RAM connectors could be the difference between a great keyboard and an unusable one, or a mediocre battery and a great one. DIMMs and SO-DIMMs are great for their toolless installation and ease of use, but if they're too large to fit a design, those advantages are meaningless. And IMO a RAM upgrade is a sufficiently advanced and infrequent thing that requiring a screwdriver is ... well, not an issue. Thus I think this is a good initiative from Dell even if it is likely largely motivated by profits - if nothing else, it can serve as a showcase of how a more up-to-date memory connector might look.